24. CLINICIAN CONTEXT MATTERS: FAMILIAL GENETIC RISK AND VARIATION IN DEPRESSION DIAGNOSIS ACROSS GENERAL PRACTITIONERS

Background Primary care physicians are the main gatekeepers to depression diagnosis and treatment, yet they differ in the diagnostic thresholds they apply. If case status depends partly on which clinician a patient sees, this has implications for public health, service planning, and the interpretation of depression phenotypes in genetic and family‑based research. We aimed to quantify variation in depression diagnosis across general practitioners (GPs) in Norway and to test whether familial risk for depression operates both at the individual and practice (GP caseload) level. Methods In this nationwide cohort study, we exploit primary health care registry data and extended population-wide pedigrees to identify depression cases (ICPC-2) and controls and compute family genetic risk for depression (FGRS). The sample includes 4,846,923 adults with at least one GP contact between 2006 and 2026, contributing 34,545,184 person‑years across approximately 11,000 GPs. For each GP, we calculated a caseload FGRS as the mean FGRS of their patients, and decomposed patient FGRS into a within‑GP component (individual deviation from that GP’s mean) and a between‑GP component (practice‑level caseload FGRS). The primary outcome was annual depression diagnosis in primary care. We fitted multilevel logistic regression models with random GP intercepts to estimate between‑GP variance and the individual and contextual effects of FGRS. Results There was substantial between‑GP variation in depression diagnosis that persisted after adjustment for patient sex, calendar year, and individual familial risk. Higher patient FGRS was associated with increased odds of receiving a depression diagnosis within the same GP. In addition, patients with similar individual FGRS had meaningfully different probabilities of diagnosis depending on whether they were seen in practices with low‑ versus high average patient depression risk. This between-GP variance is attenuated when adjusting for the average FGRS at the practice, but not eliminated. Ongoing analyses extend these models to examine the role of patient–provider matching, leveraging FGRS computed for GPs themselves and the subset of GPs who also appear as patients. Discussion Norwegian GPs shape how familial liability to depression is translated into recorded diagnoses. Depression diagnoses derived from primary care registries are therefore partly context‑dependent phenotypes, with important implications for interpreting familial and genetic associations in genomic studies.